What is agentic commerce and how it changes your online store

Agentic commerce is redefining the way we consume products and services online. It is a form of shopping in which an artificial intelligence (AI) agent does the heavy lifting for the customer: it interprets the need, searches for products, compares options, checks inventory, selects the store, and, if properly authorized, completes the payment. For the online retail sector, this represents a paradigm shift: a significant portion of traffic will no longer consist of humans browsing visual interfaces, but rather software processing structured data. With tech giants like Google, OpenAI, Visa, and Shopify leading this infrastructure, agentic commerce is moving from science fiction to a rapidly expanding market standard.

What is agentic commerce, in practice

An AI agent differs from a traditional virtual assistant (like a basic chatbot) because it does not just provide answers or lists of links, but acts on behalf of the user. If a consumer asks, "find me a pair of running shoes under 500 reais that arrives before Friday," a classic assistant would point to various stores. The AI agent goes further: it cross-references data, eliminates options without the desired size or with slow shipping, chooses the best offer, and, upon user approval, makes the purchase.

For all of this to happen in an integrated and secure manner with real money, the infrastructure requires:

• Precise comprehension: The agent needs to read product data and specifications without ambiguity.
• Identification and Security: The store needs to know that whoever is accessing its site is a trusted authorized agent, not a malicious scraping bot.
• Integrated payment: The payment authorization must be clear, protected, and subject to limits defined by the user.

How it differs from traditional e-commerce

In classic e-commerce, the main battle of digital marketing is to position the site on the first page of search engines like Google. The sales funnel depends on clicks, browsing the store, adding products to the cart, and filling in payment details by the consumer themselves.

In the agentic commerce scenario, the dynamics change significantly:

• The battle starts earlier: The first barrier is ensuring the AI knows and recommends your store when a user asks about the product. If the AI does not mention your brand, the agent will never reach your page.
• Data is the new storefront: Instead of focusing only on page design (UI/UX), stores need to have perfect structured data (price, inventory, shipping, and return policies) that machines can read (such as Schema.org via API).
• Hyper-qualified traffic: Direct human traffic may decrease, but AI agents will arrive with the purchase decision practically made.

Context and background: the path to 2026

Agentic commerce did not appear overnight; it is the result of the accelerated evolution of generative AI tools and digital payment infrastructure. Recent milestones that have consolidated this trend include:

• September 2025: OpenAI launched Instant Checkout, allowing purchases directly within ChatGPT (initially in the US). In partnership with Stripe, the Agentic Commerce Protocol (ACP) was published as open source, establishing that results should be organic and relevance-based.
• October 2025: Cloudflare, Visa, and Mastercard created security standards to distinguish authorized agents from harmful bots. Agents started using public keys to sign requests, ensuring security for online stores.
• January 2026: Google introduced the Universal Commerce Protocol (UCP), developed in collaboration with major players like Shopify, Etsy, Target, and Walmart, alongside support from payment giants (Stripe, Adyen, Visa, Mastercard). This protocol enabled purchases in the AI Mode of Google Search and the Gemini app.
• March 2026: OpenAI redesigned its strategy. Instead of focusing only on the internal ChatGPT checkout, it shifted payment to each store's apps/sites. Conversely, it focused on product discovery via chat. Shopify followed the move, integrating its products natively into ChatGPT searches.
• May 2026: Google launched the "universal cart" in the US, integrating purchases from Search and Gemini, using a version of the payment protocol (AP2) that enables autonomous purchases by agents while respecting user-pre-approved limits.
• September/October 2026 (Personal Agents): The market exploded with autonomous launches. Meta launched Muse; OpenAI, Dots; tools like Grok Bot (using Stripe Link), Instinct, and Tab emerged quickly, making purchases directly on store websites. While platforms like Shopify opened their doors to these agents, giants like Amazon initially opted to block them.

The numbers driving the shift

Market estimates, although based on prospective scenarios, demonstrate the magnitude of what is to come:

• 310 billion Euros: This is the forecast for the value of online purchases in Europe that could be brokered by AI agents over the next decade (Sopra Steria, 2026).
• 10% to 20%: The estimated proportion of e-commerce in the US that could be conducted via AI agents by 2030 (Morgan Stanley Research, 2025).

Despite the optimism, data also shows that absolute trust in autonomous purchasing (where AI decides everything) remains cautious; most users (about 50% in surveyed countries) prefer the AI agent to curate options, delegating the final click decision to a human. The current goal is to be on the agent's "shortlist" of recommendations.

What changes for online store operations

For a brand to be found, chosen, and purchased by an AI agent, it needs to overcome three structural stages:

1. Technical Access: The first visitor to your store might be a robot (an agent). If anti-bot policies (or the robots.txt file) block all automated access, the agent will never see the store. Reliable identification has become crucial.
2. Structured Data: The "front-end" does not matter to the machine. Product information (price updated via API, real-time inventory, clear technical specifications, machine-readable return policies) must be impeccable.
3. AI Discovery: Before the visit occurs, the AI must "know" the store exists and consider it relevant. Being well-indexed in language models has become the new search optimization (the new SEO).

Practical Tip for shop owners: Test your brand's visibility today by asking ChatGPT or Google Gemini about your product category, just as a customer would, and see if your store appears on the "shortlist."

What to expect next

In the coming years, protocol standards are expected to unify or establish strong interoperability bridges. The UCP (Google) and the ACP (OpenAI/Stripe) are likely to be massively adopted by e-commerce platforms, standardizing how products are exposed to agents.

We will also see the evolution of the "Agentic Economy"—where different agents talk to each other, negotiating shipping costs or deadlines in milliseconds before presenting final options to the human consumer. In the Latin American market, stores that quickly adopt clear data integrations (APIs and structured product feeds) will get ahead of competitors, as the technical entry barrier will be a strict filter for those slow to adapt.

Frequently Asked Questions

Will agentic commerce end traditional online stores? No. Agentic commerce will act as a new transaction layer. Consumers will still access traditional sites for certain purchases (like complex clothing or luxury goods), while using agents for recurring purchases or those focused on specific specs and price.

How can I ensure my store is found by AI agents? Ensure your product data is formatted in a machine-readable way (using Schema.org, for example), provide APIs with real-time prices and inventory, and do not block authorized agents with generic anti-bot rules or robots.txt.

Does the customer fully trust agents to make payments? At the moment, the main trend is "assisted" agents, where the AI searches, filters options, and even fills the cart, but the final payment authorization requires confirmation from the human customer. The level of autonomy will increase progressively as the technology consolidates robust security and anti-fraud mechanisms.